The present invention relates to management of virtualized workloads, and more specifically, to cache mobility. In a dynamic computing environment (e.g., cloud computing environment), applications being executed may be implemented on resources that are not co-located or may be moved from one physical system to another. For example, when a customer of a cloud computing environment executes an application, the application may be implemented on one cloud resource that accesses information from a database stored on a different cloud resource that is separated by several ethernet switches. As another example, unbeknownst to a cloud customer, an application being executed on one cloud resource may be moved to another (physical) cloud resource for any number of reasons. The reasons may include a crash, upgrade, maintenance, or management decision (e.g., consolidating several users onto one resource rather than spreading them out to increase power usage efficiency).